
espnet/owls_4B_180K
Automatic Speech Recognition
•
Updated
•
163
•
4
🦉 A suite of Whisper-style models from 250M to 18B parameters. Trained on up to 360K hours of data. 16k sampling rate.